JPH0428679B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Technical Field The present invention relates to an InP single crystal with a low dislocation density and a method for manufacturing the same.

InP single crystals have wide applications as substrates for light-emitting diodes, laser diodes, photodiodes, transistors, and ICs.

(a) Prior art InP single crystals can be produced using the LEC method (liquid capsule method) or
It can be made using the HB method (horizontal Bridgeman method).

In the LEC method, raw materials and liquid capsules are placed in a crucible and heated several dozen times with inert gas in a pressure-resistant container.
By applying ATM pressure, the raw material is melted, and a single crystal is pulled from the melt using a seed crystal.

When the raw material and liquid capsule are melted by a heater, the liquid capsule covers the upper part of the raw material melt. Since the inert gas applies high pressure on top of this, it is possible to prevent P from escaping.

The single crystal pulled by the LEC method is cylindrical, and has the advantage that when sliced into wafers, circular wafers can be obtained immediately. Another advantage is that since a boat is not used, there is no silicon contamination and it is easy to obtain the desired electrical characteristics.

(c) Problems with conventional technology InP single crystals pulled by the LEC method often have many crystal defects due to thermal strain during growth.

There are several types of crystal defects, and their degree can be evaluated by etching pit density (EPD). This is because when a single crystal is sliced into wafers and etched with an appropriate etching solution,
Etchipits appearing on the surface are counted under a microscope and converted to the number per unit area.

The etching pit corresponds to some kind of crystal defect. When devices such as light-emitting elements, light-receiving elements, and transistors are fabricated on InP crystals, crystal defects degrade the characteristics of these devices and shorten their lifetimes. Also, if there are variations in the etchipit density,
When making FETs, the pinch-off voltage etc. will vary.

In this way, it is desirable that the etching pit density be as low as possible, and moreover, it is better to have less variation.

Thus, it is strongly desired to produce InP single crystals with low dislocation density.

(d) Impurity hardening Pure metals are generally soft and brittle, but the introduction of impurities often makes them harder and stronger.

If crystal defects are caused by a weak lattice structure, then by adding impurity elements,
It should be possible to strengthen InP single crystals.

In fact, by adding more than ¹⁰¹⁸ atoms/ ^cm3 of Zn and S as impurities to InP, the etch pit density can be increased.
The phenomenon that EPD decreases is widely known.

However, Zn and S are elements of groups, respectively, and are not elements of group . Zn, S is
Substitution of In or P atoms results in free electron or hole charge carriers in the crystal. Then, it becomes an n-type or p-type semiconductor. Such materials cannot be used as substrates for devices that require semi-insulating crystals, such as field-effect transistors (FETs).

The main uses of InP substrates are light emitting devices for optical communications,
Or it is a substrate of a light receiving element. These are n-type, p-type
It does not matter if it is a type wafer. however,
When Zn or S is added, for example, 1×10 ¹⁸ cm ^{-3 or} more, the number of carriers increases, so the light absorption coefficient increases and the light extraction efficiency from the substrate side (in the case of a light emitting device)
It also has the disadvantage that it decreases. However, Zn,
When the concentration of S is 3×10 ¹⁸ cm ^-3 or more, impurity effects due to these can be expected.

The present invention does not contain Zn, S or 1×
It is effective when the concentration is below 10 ¹⁸ cm ^-3 and no impurity effect occurs.

(e) Isoelectronic impurity hardening In InP crystal, In is a group element and P is a group element. Suppose that a group or group element is added as an impurity into InP. If the group element replaces the In site and the group element replaces the P site, then
These additions should not change the electrical properties. Since the group elements have different covalent bond radii and different numbers of electrons, the above-mentioned substitution between homologous elements is actually likely to occur.

Group elements other than In include B, Al, and Ga.

Group elements other than P include N, As, and Sb.

For example, suppose InP is doped with Ga. It is known that the segregation coefficient of Ga in InP is about 5.

The distribution coefficient refers to the proportionality constant of the concentration of a certain impurity, which has a proportional relationship between the solid phase and the liquid phase when the solid phase and the liquid phase are in a state of thermal equilibrium. The concentration in the solid phase (wt%) is
The segregation coefficient is divided by the concentration in the liquid phase (wt%).

Putting an impurity with a segregation coefficient k into the raw material melt,
When a crystal is produced, and the solidification rate is g, the impurity concentration c changes in the crystal according to the formula: c=C ₀ k (1-g) ^k-1 (1). c ₀ is the initial concentration of impurities in the melt.

This formula can be derived as follows. Let L ₀ be the total weight of the initial raw material melt. When the solidification rate is g, the amount of melt is L ₀ (1-g), and the amount of solidification is
L ₀ g. The solidified minute amount is expressed as d (L ₀ g). The amount of impurities contained in this is cd(L ₀ g) (2).

On the other hand, the impurity concentration in the melt is expressed as c/k. The amount of impurities in the melt is cL ₀ (1-g)/k (3). The derivative of this gives the amount of reduction of impurities in the melt. Therefore, the sum of this differential and equation (2) is equal to 0.

cd(L ₀ g)+d(cL ₀ (1-g)/k)=0 (4) Solve this to obtain equation (1).

G. Jacob et al., J. Crystal Growth 61 (1983)
p.417−424 “DISLOCATION−FREE GaAs
AND InP CRYSTAL BY ISOELECTRONIC
In ``DOPING'', we report on an experiment in which InP crystals were doped with one type of isoelectronic impurity among Sb, As, and Ga and pulled using the LEC method.

It weighs about 50g and is small with a diameter of 15mm to 20mm.
It is an InP single crystal grown in the <111> direction. All have carrier concentrations of 10 ¹⁶ /cm ³ or higher and are reported to be n-type crystals. It states that the electrical properties are the same as non-doped InP.

The undoped InP is said to be entirely single crystal, with an etching pit density EPD of 10,000 to 10/cm ² .

When InSb is added to the melt at a concentration of 3.6wt% relative to InP, the upper 4/5 becomes a single crystal.
The EPD was said to be several 10 ² /cm ² .

When InSb was added to the melt at a concentration of 7.6 wt% compared to InP, the EPD was ~10 ² /cm ² and twin crystals were formed in the upper half.

When InAs is added to the melt at a concentration of 6.2wt% relative to InP, the upper 3/4 becomes a single crystal.
The EPD was said to be several 10 ³ /cm ² .

When GaP was added to the melt at a concentration of 2.3 wt% compared to InP, the entire material became a single crystal, with no dislocations at all in the upper part.

According to Jacob's results, when we introduce isoelectronic impurities,
It can be seen that it is effective in reducing the EPD of InP. However, it has the disadvantage that it often does not become a single crystal.

Jacob is a small, light crystal that is pulled up, but industrially it has a diameter of 2 inches (approx. 50 mmφ).
However, it is meaningless unless a single crystal weighing several kilograms can be obtained.

Since the LEC method pulls the crystal upward, cooling progresses in the vertical and radial directions. As the radius increases, the thermal strain increases, so the difficulties increase, both in terms of single crystallinity and EPD.

Jacob's method has the following characteristics when introducing impurities.

(1) Impurities are not added as single elements, but in the form of compounds. The other side is the component element of the host crystal. In other words, the initial numbers of group and group elements in the raw material melt are equal.

(2) Only one type of impurity can be added.

(f) Discussion It is easy to understand that if InP is doped with a large amount of impurities such as Sb or As, it will not become a single crystal in the lower half.

These have a segregation coefficient k smaller than 1 in InP. Therefore, as the pulling of the single crystal progresses, the impurities become concentrated in the melt, and the impurity concentration increases to such an extent that the single crystal cannot be maintained.

When InP is doped with Ga, the segregation coefficient is approximately 5, which is greater than 1, so Ga is incorporated into the solid at the beginning of pulling. The concentration becomes thinner rapidly. For this reason, it seems that the EPD suppressing effect only exists in the upper part of the crystal.

There is something important that is overlooked by Jacob.

This means matching the bond lengths between atoms. - Group compounds all have the same crystal structure.

In a pure InP crystal, the In-P bond length is
It is 2.54 Å. The bond length of GaP crystal is 2.36 Å. The In-Sb bond length of InSb crystal is 2.8 Å.
The In-As bond length of InAs crystal is 2.62 Å.

When doped with impurities that make the bond length longer than that of the host crystal InP, compressive stress will occur in the crystal. The opposite may be the case.

This stress increases thermal strain. This inventor thinks so.

When InP is used as the host crystal, the In-As bond length is 3.1% longer. The bond length of GaP is 7.1% shorter. InSb
The bond length of is 10.2% longer.

Due to its single crystallinity and low EPD properties, the present inventor
I think lattice matching should be achieved. In other words, by combining impurities with long bond lengths and impurities with short bond lengths, the average value of the bond length deviations is made close to 0.

For example, if the ratio of As to Ga can be set to about 7:3 in the crystal, the average bond length can be equal to the bond length of In-P photo elements.

Furthermore, since the segregation coefficient of impurities is not 1, impurities cannot be uniformly distributed throughout the single crystal. Those with segregation coefficients less than 1 are concentrated in the tail, and those greater than 1 are concentrated in the head.

Therefore, impurities with segregation coefficients smaller than 1 and impurities with a segregation coefficient of 1
By combining larger impurities, it is possible to distribute the impurities more evenly from the head to the tail of the single crystal.

The segregation coefficient of Ga is larger than 1, and that of As and Sb is smaller than 1.

When it comes to combining a material with a segregation coefficient larger than 1 and a material with a smaller segregation coefficient, Ga and As may be used, or Ga and Sb may be used.

However, from the point of view of lattice matching mentioned earlier, it is clear that it is better to combine Ga and As.

(g) Structure of the invention In order to reduce the EDP of InP crystal, Ga or As
It is known that doping is effective. The upper and lower limits when added alone are also known.

This cannot necessarily be said to be the limit in terms of EPD reduction.

(1) Regarding the concentration of Ga in InP, when the Ga concentration in InP is increased, Ga precipitates as a single substance rather than as a compound. The precipitation limit is about 3×10 ²⁰ atoms/cm ³ in terms of concentration in the crystal.

By adding Ga to InP, impurity hardening occurs at 10 ¹⁸ atoms/cm ³ or more. This is not a clear limit, but indicates that there is no effect below this limit.

As a result, it is found that a preferable value for the Ga concentration in crystalline InP is 10 ¹⁸ to 3×10 ²⁰ atoms/cm ³ . This is tentatively called the optimum concentration of Ga.

(2) About the concentration of Ag in InP The precipitation limit of As in InP is 5×10 ²⁰ atoms/cm ³
That's about it. The impurity hardening effect of As is
It is considered to be ¹⁰¹⁸ atoms/cm ³ or more. Ultimately, the preferred concentration of As in the InP crystal is 10 ¹⁸ to 5×10 ²⁰ atoms/cm ³ . This is called the preferred concentration of As.

Therefore, the InP single crystal of the present invention (1) is doped with both Ga and As as impurities, and (2) is doped with either Ga or As or both at any point in the single crystal. It can be characterized by being at a suitable concentration.

Since k is different from 1, some measures are required to do this.

The upper limit of the preferable concentration is b and the lower limit is a. From equation (1), when k is smaller than 1, the impurity concentration c is minimum when the solidification rate g is 0, and as s increases, c increases.

When k is greater than 1, c is maximum when g is 0,
As g increases, c decreases.

The condition (2) above is for Ga or As, a≦c≦b (5) That's what it means.

Condition (5) may be satisfied for Ga and As at the same time, or only one of them may be satisfied. However, c must not exceed the precipitation limit. c may exceed the lower limit a. There is such an asymmetry regarding the upper and lower limits.

To avoid confusion, I will add a sign here.

Regarding As, the segregation coefficient is m (0.4), the initial melt concentration is x, the solidification rate is g, and the concentration in the crystal is u. Instead of equation (1), we can write As: u=xm/(1-g) ^1-m (6). As mentioned earlier, in the crystal
Letting the lower limit of the preferable concentration of As be E and the upper limit F, E=1×10 ¹⁸ atoms/cm ³ (7) F=5×10 ²⁰ atoms/cm ³ (8).

Regarding Ga, the segregation coefficient is n(5), the initial melt concentration is y, the concentration in the crystal is v, and the solidification rate is g. (1)
Instead of the equation, we can write Ga:v=yn(1-g) ^n-1 (9). Let H be the lower limit and K be the upper limit of the preferable concentration of Ga in the crystal.

H=1×10 ¹⁸ atoms/cm ³ (10) K=3×10 ²⁰ atoms/cm ³ (11).

u and v must not exceed the upper limit. However, there is no problem with the lower limit as long as only one of them exceeds it.

The conditions for InP of the present invention are: (i) u≦F (12) (ii) v≦K (13) (iii) E≦u or H≦v (14) This means that the following conditions are satisfied.

Of these, (13) is clear, but (12) and (14) are not necessarily clear.

With respect to g, v is a downward-sloping curve, and u is a downward-sloping curve to the right.

(13) and the condition that both Ga and As are doped, it can be seen that H/n≦y≦K/n (15). This means that the initial concentration y of Ga in the melt is between 2×10 ¹⁷ atoms/cm ³ and 6×10 ¹⁹ atoms/cm ³ .

When the initial melt concentration y of Ga is given, the maximum solidification rate G(y) that satisfies equation (14) is found.

It is.

For As, when the solidification rate g is G(y), u=
The initial concentration such that E gives the lower limit of x. In other words, the initial melt concentration x of As is x≧E/m(1−G(y)) ^1−m (17).

The upper limit of x depends on the maximum value Gs of the solidification rate g.
Conditions (12) to (14) of the present invention when the solidification rate g is from 0 to Gs.
Suppose that is satisfied.

When g is equalized to Gs, the upper limit of x is u=F, so it is given by x≦F/m(1−Gs) ^1−m (18). Substituting (16) into (17),
Writing the conditions regarding y and x together, H/n≦y≦K/n (19) Substituting m=0.4, n=5, and other numerical values of E, F, H, and K, 2×10 ¹⁷ ≦y≦6×10 ¹⁹ (21) 2.77×10 ¹⁸ (10 ¹⁷ /y) ^0.15 ≦x≦ 1.25×10 ²¹ (1−
Gs) ^0.6 (22).

The above conditions are the gist of the present invention. What has already been explained using formulas will be explained again using diagrams.

FIG. 1 is a diagram showing the relationship among the concentrations u and v of As and Ga in the InP crystal of the present invention, the initial melt concentration x and y, and the upper limit and lower limit solidification rate g.

In (a), g is plotted on the horizontal axis, and x is the parameter.

It is a graph showing the concentration u of As in crystals. The upper limit is F and the lower limit is E. Since u diverges when g=1, we will cut it at the critical solidification rate Gs. Many u-lines can be drawn using x as a parameter. u is
Since it follows equation (6), it is a curved line, but for simplicity it is shown as a straight line. The graph of u is simply called the u line.

(b) is a graph showing the concentration v of Ga in the crystal with y as a parameter. This is a graph of equation (9), but it is shown as a straight line for simplicity. Upper limit is K, lower limit is H
It is. The intersection of the v line passing through K with the lower limit line HH is
Goalkeeper. Since the segregation coefficient (n=5) is too large than 1, the v-line does not fall into the preferred region unless g is less than or equal to Gk. (c) is the concentration u of As and Ga in the crystal,
v are shown combined into the same graph.

The lower limit and upper limit are different for v and u, but for ease of understanding, they are shown at the same position on the left and right. The value of yn for v is between H and K. The larger the value of yn, the larger the range of existence of x, which is a parameter of u. When the v line passes through point K, the solidification rate at the intersection with the lower limit line is defined as Gk. u line passing through Gk and g=
Let Wk be the intersection of Gs. In the limit where yn is in K,
The u-ray lies between wide values of H Gk Wk F.

Generally, the solidification rate at the point where the v line intersects the lower limit H is defined as G
(y). Let W be the intersection of line u passing through G(y) and g=Gs. The existence range of u-rays is limited to HG(y)WF.

The limit at the intersection of the u line with g=Gs is (20)
It corresponds to the formula.

Now, equations (19) and (20) determine the range of values of the initial melt concentrations x and y of As and Ga. Figure 1d illustrates this.

If you want to define this using the As and Ga concentrations in the crystal, give x and y and obtain the ranges of u and v from (6) and (9).

However, the range of values of u and v is a function of the solidification rate g.

The solidification rate g can also be considered to be coordinates corresponding to the distance from the head of the crystal to the tail. g=0
is the head and g=Gs is the tail. Gs is the value obtained by dividing the weight of the grown single crystal by the amount of raw material charge.

The As and Ga concentrations in the InP crystal at the head are mx,
It is ny. The range of this value can be easily determined from equations (19) and (20). When mx and ny are expressed as U and V, H≦V≦K (23) becomes. The As and Ga concentrations U and V in the InP crystal at the head are determined by substituting constants: 1×10 ¹⁸ ≦V≦3×10 ²⁰ (25) 10 ¹⁸ × (10 ¹⁸ /V) ^0.15 ≦ U≦5×10 ²⁰ (1-Gs) ^0.
Get ⁶ (26).

Ga concentration V at the head of InP single crystal is 1×10 ¹⁸
atoms/cm ³ or more, and the As concentration U is expressed by equation (26).

(H) Conversion Up until the previous section, we have used atoms/cm ³ to express concentration.

Other methods of expressing concentration include atomic percent (at%),
Weight% (wt%), etc. Conversion between these is shown below.

The density of InP is ρ, the molecular weight is M, and Avogadro's number is
Assuming L ₀ , there are ρL ₀ /M (27) of In or P atoms per 1 cm ³ . Since there is one In and one P,
Multiply equation (27) by 2 to obtain the number of host atoms per 1 cm ³ . The atomic ratio of impurities to host atoms is given by (number of atoms)=M/2ρL ₀ ×(atoms/cm ³ ) (28). ρ=4.787g/ ^cm3 , M=145.8, _L0
= 6.02×10 ²³ , so (atomic ratio) = 2.53×10 ^-23 × (atoms/cm ³ ) (29). Multiply this by 100 to get at%.

Instead of (25) and (26), if we express the Ga concentration V and As concentration U in at%, then 0.0025≦V(at%)≦0.76 (30) 0.001/V ^0.15 ≦U(at%)≦1.26( 1-Gs) ^0.6 (31). To obtain x and y, divide U and V by m and n.

Also, when expressed in weight%, the molecular weight is InP=
145.8, As=74.9, and Ga=69.7, so 0.0024≦V (wt%)≦0.73 (32) 0.001/V ^0.15 ≦U (wt%)≦1.29 (1−Gs) ^0.6 (33). To get x, y, divide U, V by m, n.

0.00048≦y (wt%)≦0.145 (34) 0.002/y ^0.15 ≦x (wt%)≦3.23 (1−Gs) ^0.6 (35).

(k) Pulling device The present invention can be used for both the LEC method and the HB method. These devices are known. A known device for the LEC method will be briefly explained.

FIG. 2 is a sectional view of the LEC lifting device.

1 is a crucible, and 2 is a susceptor that supports the crucible 1. A raw material melt 3 is contained in a crucible 1, and a liquid capsule 4 covers the melt.

A lower shaft 8 supports the susceptor 2 so that it can rotate and move up and down. A seed crystal 6 is attached to the lower end of the upper shaft 7. The upper shaft 7 is also rotatable and can be moved up and down. The seed crystal 6 is immersed in the raw material melt 3, and the single crystal 5 is pulled up. A heater 9 heats the raw material melt 3. The entire container is covered with a pressure-resistant container and filled with inert gas or nitrogen gas. The gas pressure is several tens of atm, holding down the liquid capsule 4,
Prevents volatilization of group elements.

(J) Example Using the LEC lifting device as shown in Figure 2,
Ga, As-doped InP crystals were grown. The growth direction is the <100> direction.

Raw materials are InP polycrystal 1000g (6.86mol) Liquid capsule B ₂ O ₃ 150g Impurity-added InAs 6300mg (0.0332mol) GaAs 500mg (0.00346mol) Sn 2100mg (0.0177mol) As is 0.27wt%, Ga is 0.024wt% .

This is U=0.95×10 ²⁰ cm ^-3 , V=0.99×10 ¹⁹ cm ^-3
That's what it means.

These raw materials were placed in a quartz crucible and pulled under the following conditions.

Melting temperature: 1100° C. _N2 gas atmosphere: 49 atm Pulling speed: 7 mm/H Upper shaft rotation speed: 3 to 20 rpm Lower shaft rotation speed: 5 to 30 rpm As a result, a single crystal with a diameter of 2 inches was pulled. We were able to convert 92% of the raw material weight into single crystals.

This ingot was sliced into wafers having (100) planes. Etch this and EPD
was measured.

FIG. 3 is a graph showing the EPD measurement results.
The horizontal axis is the distance (mm) from the wafer center. The vertical axis is EPD (pieces/cm ² ).

For the above ingot, the average EPD was 7000 pieces/cm ² for the head and 5000 pieces/cm ² for the tail, excluding the peripheral 5 mm. The EPD distribution is W-shaped in the plane. Although it is high at the center and periphery, there is a region in the middle where the EPD is almost 0. Further, no precipitation of Ga or As was observed throughout the ingot.

Single crystal EPD with no Ga or As added to the same graph
The measurement results are shown. EPD is on the order of 20,000 to 40,000/cm ² . It can be seen that the EPD is significantly reduced compared to this. This single crystal has Sn concentration,
The InP charge amount is the same, and the only difference is that As and Ga are not doped.

(S) Impurity addition, raw material InP The methods of adding Ga to InP include Ga alone,
It is added in the form of GaAs, GaP, etc.

Methods for adding As to InP include As alone,
Added as GaAs or InAs.

Sn can be added alone.

The raw material for InP may be directly synthesized from In and P, or polycrystalline InP may be used.

Furthermore, polycrystalline InP containing As or Ga impurities of 1×10 ¹⁷ atoms/cm ³ or more may be used as the starting material.

(C) Growth method The present invention can be applied to either the LEC method or the HB (horizontal Bridgeman method). Head (g=0) and tail (g
= Gs), in any case, the first part to solidify is the head, and the last part to solidify is the tail.

(S) Addition of impurities other than isoelectronic impurities In addition to isoelectronic impurities, it is also possible to add the following impurities.

(1) Semi-insulating crystal containing impurities such as Fe, Cr, and Co.

(2) P-type crystal containing group elements such as Zn and Cd.

(3) N-type crystal containing group elements such as S, Se, and Te.

(4) P- or N-type crystals containing Si, Ge, Sn, etc. elements.

(C) Effects (1) InP single crystal has fewer lattice defects. EPD is low when made into wafers.

Since Ga and As do not precipitate and have complementary impurity hardening effects, the EPD is low over the entire length.

(2) The amount of addition is smaller than the usual addition of only Ga or only As. For this reason, there is no precipitation, and the entire material tends to become a single crystal. Since the head part is made of Ga (n=5) with a large segregation coefficient and the tail part is made of As with a small segregation coefficient (m=0.4) to exert an impurity hardening effect, the mixing amount is small.

(3) Since both Ga and As are isoelectronic impurities,
The electrical properties of the InP single crystal remain unchanged.

(4) The Ga-P bond length is shorter than the In-P bond length, and the In-As bond length is longer than the In-P bond length. For this reason, when Ga and As are added as impurities, the deviations in the lattice dimensions mutually compensate and the stress is canceled out. For this reason, it tends to become a single crystal and has the effect of reducing EPD.

[Brief explanation of drawings]

FIG. 1 is a diagram for explaining the ranges of the concentrations u and v of As and Ga in the crystal and the initial melt concentrations x and y that should be added to InP in the method of the present invention. (a) is a function of the solidification rate g in InP crystals.
The concentration u of As in the crystal is shown. Many lines can be drawn using the initial concentration x as a parameter. E is u
is the lower limit, and F is the upper limit. (b) shows the concentration v of Ga in the crystal as a function of the solidification rate g in the InP crystal.
Many lines can be drawn using the initial concentration y as a parameter. H is the lower limit of v, and K is the upper limit. (c) is
The concentration diagram of As and Ga is shown on one graph, and the upper limits K and F and the lower limits H and E are matched. Gs is the solidification rate that gives the limit of single crystal (no precipitation). Gk is the solidification rate at the lower limit H when the v line passes through point K. Parameter y
When is given as a value between H/n and K/n,
Let G(y) be the solidification rate at the lower limit of the v line. G
Draw a line u through (y) and let W be the intersection with the critical solidification rate Gs. existence of u or within HGWF. The parameter x can be found from here. (d) is a graph showing the existence range of As and Ga concentrations x and y in the initial melt. FIG. 2 is a sectional view of a known LEC lifting device.
FIG. 3 is a graph showing the EPD measurement results of InP wafer in Example. The horizontal axis is the distance (mm) from the wafer center, and the vertical axis is the EPD (pieces/cm ² ). The lower curve is Ga, As-doped InP (Sn), and the upper curve is Ga, As-doped InP (Sn).
EPD of InP(Sn) crystal without As doping is shown. 1...Rutsubu, 2...Susceptor, 3...Material melt, 4...Liquid capsule, 5...Single crystal, 6...
Seed crystal, 7...upper shaft, 8...lower shaft, 9...heater.

Claims (1)

[Claims] 1 Contains Ga and As as impurities, and the concentration U of As and the concentration V of Ga at the crystal head are 1×10 ¹⁸ ≦V, where Gs is the solidification rate of the crystal tail. ≦3×10 ²⁰ 10 ¹⁸ × (10 ¹⁸ /V) ^0.15 ≦U≦5×10 ²⁰ (1-Gs) ^0.
InP single crystal with a low dislocation density, characterized by a dislocation density in the range of ⁶ . 2. An InP single crystal with a low dislocation density as described in item 1 of the claimed scope, which contains Si, Ge, and Sn as impurities. 3. The low dislocation density InP single crystal according to claim 1, which contains Fe, Cr, or Co as an impurity. 4. The low dislocation density InP single crystal according to claim 1, which contains Zn or Cd as an impurity. 5. The low dislocation density InP single crystal according to claim 1, which contains S, Se, or Te as an impurity. 6 As and Ga are added as impurities to InP polycrystal, and the initial concentrations x, y of As and Ga in the raw material melt are
However, when Gs is the solidification rate of the crystal tail, 2×10 ¹⁷ ≦y≦6×10 ¹⁹ 2.77×10 ¹⁸ (10 ¹⁷ /y) ^0.15 ≦x≦1.25×10 ²¹ (1−
Gs) is in the range of ^0.6 , and a single crystal is grown from this raw material melt.
Manufacturing method of InP single crystal. 7. InP single crystal with low dislocation density according to claim 6, wherein the InP polycrystal is a raw material containing 1×10 ¹⁷ atoms/cm ³ or more of Ga or 1×10 ¹⁷ atoms/cm ³ or more of As. manufacturing method. 8 Ga is added in the form of Ga alone or a compound of GaAs or GaP, and As is added in the form of As alone or as a compound of GaAs or GaP.
A method for producing an InP single crystal with a low dislocation density according to claim 6 or 7, wherein GaAs and InAs are added in the form of a compound.